The Infona portal uses cookies, i.e. strings of text saved by a browser on the user's device. The portal can access those files and use them to remember the user's data, such as their chosen settings (screen view, interface language, etc.), or their login data. By using the Infona portal the user accepts automatic saving and using this information for portal operation purposes. More information on the subject can be found in the Privacy Policy and Terms of Service. By closing this window the user confirms that they have read the information on cookie usage, and they accept the privacy policy and the way cookies are used by the portal. You can change the cookie settings in your browser.
This paper proposes a low-cost video-based Real-Time Pupil-Tracking embedded system which will allow people with reduced mobility to control a wheelchair through their eyes. The main aspect of the method is its capacity to be implemented in a portable computing system, reduced both in computing power and in RAM memory. The Pupil-Tracking system is based on Feedforward Neural Networks-using offline...
Recognition is always an interesting aspect of visual processing, especially for systems that requires intuitive perception like robotics or human-machine interactions. In this work, a color recognition system based on Evidence Theory is applied for a scenario of the NAO robot that recognizes the color of a requested ball. The robot employs multi-cameras to reduce uncertainties, and the Dempster-Shafer...
Though the classical robotics is highly proficient in accomplishing a lot of complex tasks, still it is far from exhibiting the human-like natural intelligence in terms of flexibility and reliability to work in dynamic scenarios. In order to render these qualities in the robots, reinforcement learning could prove to be quite effective. By employing learning based training provided by reinforcement...
Despite significant recent progress, the best available computer vision algorithms still lag far behind human capabilities, even for recognizing individual discrete objects under various poses, illuminations, and backgrounds. Here we present a new approach to using object pose information to improve deep network learning. While existing large-scale datasets, e.g. ImageNet, do not have pose information,...
Traffic prediction systems are currently the most important techniques, as they can be wildly applied in different aspects. Given a set of past traffic data, a traffic prediction system is able to predict the future traffic conditions. However, the existing traffic prediction systems are hard to implement and are quite expensive. Hence, this work proposed a Matlab-based traffic prediction system,...
When processing video, it is normally assumed that cameras are vertically oriented such that people appear upright, which helps simplify subsequent processing such as person detection. In real situations, due to the need to provide maximum coverage of the viewing space, cameras are usually placed with arbitrary orientations so the apparent vertical axis of the videos captured may not correspond to...
The trajectory tracking system of particle motion on sieve surface was designed by the combination of the analysis of image sequences based on binocular stereo vision and three-dimensional position reconstruction based on artificial neural network. Firstly, the calibration plane with uniformly distributed solid circles was placed in multiple positions within the effective field of view. The images...
This paper is focused on the problem of tracking an object by the head movement of robot with two cameras simultaneously, one robot camera and one fixed external camera. The goal of using external camera is to test, how it can aid the head camera of robot when the object moves out of its field of view. The setup of system is focused on comparison of robot with and without additional camera. The tracked...
Camera calibration is necessary in machine vision application field. Calibration model has nonlinear characteristics, and establishment of mathematical model is often a complicated process, but neural network can solve the complex nonlinear problem effectively, neural network has strong nonlinear approximation ability, adaptive network parameters and fast learning. This paper presents a neurocalibration...
Research in the field of Human-Computer Interaction (HCI) has become more and more frequently in our life. These related applications not only make our lives more convenient and efficiency, but also reduce the overhead costs. Users can more naturally, quickly and intuitively convey commands through different types of HCI applications to computer. Leap Motion1 is used to capture the information of...
It is well known that ancient buildings suffer a high vulnerability to hazards, which may induce unpredictable damages. For this purpose, a main objective to be pursued concerns with the development of techniques for monitoring historical buildings and immediately alerting in case of early vulnerability warnings. This paper proposes a noninvasive Neural Network-based (NN-based) approach for Monitoring...
This paper reports on the results obtained when using Radial Basis Neural Networks to classify different objects using only colour data extracted from images captured under different lighting conditions. Each network is trained with data from a single image and then tested with data from images containing the same collection of objects, but captured under different lighting conditions. The ability...
This document describes a proof of concept for a new approach for next generation wheel alignment systems. We propose to measure the relevant angels with Kohonen self organizing networks from an image of a heavy precision camera, instead of a projecting system clamped on the wheel. This has the clear advantage, that we do not need to attach a specially designed clamp which holds on to a wheel with...
The results of design and investigation on a human gesture recognition system, based on a Kinect sensor, are presented in this paper. In the presented research, we use a Kinect device as a 3D data scanner. Therefore, the 3D coordinates are calculated directly from depth images. The system's hardware description and computation method for 3D human gesture identification are presented in this study...
We developed and implemented a system for real time automatic recognition of components from a production line using neuronal networks. The capture device (Web camera) is placed over the production line and the system can identify the type of component even if this is not in the correct position or centered. The device sends the information to a recognition system (software) which identifies the type...
Virtual Generalizing Random Access Memory Weightless Neural Networks (VG-RAM WNN) is an effective machine learning technique that offers simple implementation and fast training and test. We examined the performance of VG-RAM WNN on binocular dense stereo matching using the Middlebury Stereo Datasets. Our experimental results showed that, even without tackling occlusions and discontinuities in the...
It has been proved that acquired training is important to the development of stereopsis experience. Month-old babies already have the initial experience of invariance recognition of 3D objects. There is a slight lack of precision in the interpretation of biological vision. However, the small cost and the fast speed in calculation meet the requirements of invariance recognition, the rich visual experience...
Set the date range to filter the displayed results. You can set a starting date, ending date or both. You can enter the dates manually or choose them from the calendar.